752 MR. E. S. GOODEICH ON [NoV. 12, 



layers of a ganoid scale are not formed of true enamel, but are 

 continuations of the lower layers which overlap on to the uppei- 

 surface, the scale growing by the deposition of complete con- 

 centric lamellae. The Teleostean scale he described as growing in 

 the same way, and as derived from the Ganoid. 



At the time Williamson made these interesting observations, the 

 classification of fossil fish was in a very unsatisfactory state. 

 Allied forms were frequently widely separated, or distantly 

 related forms closely grouped together. It is not surprising, then, 

 that the systematic importance of his work should have escaped' 

 the notice of ichthyologists as well as of the author himself. 

 Now that the classification is so much better understood, we cannot 

 help being struck by the value of the evidence afibrded by his- 

 researches ; and it is mainly for the purpose of calling attention, 

 to them that this paper has been written. 



Types of Scale. 



The Placoid. Scale. — Little need be said concerning the denticles; 

 of Elasmobranch fish ; attention must, however, be drawn to a 

 few important points. True denticles are universally pi'esent in 

 the living Elasmobranchs and their extinct allies. In structure 

 they are remarkably constant, always consisting of a cap of dentine 

 enclosing a pulp-cavity from which radiate numerous canaliculi 

 (text-fig. 196). The cavity often becomes subdivided into branching- 

 canals uniting at the base, and giving oflf the dentinal tubules. 

 The base is pierced by one or more openings through which the 

 pulp can communicate with the surrounding mesoblastic tissues of 

 the dermis. The dentine cap is covered on its outer surface with 

 a layer of hard transparent enamel-like substance, the exact nature 

 of which has given rise to considerable controversy. It is either 

 true enamel deposited by the epidermis, or merely an outer 

 specialised layer of dentine (Rose, 21), or a combination of both 

 (Tomes, 22). For our present purpose we may call it enamel,, 

 allowing that Tomes' view is probably correct. 



The important thing to notice is that from its earliest appearance 

 in development onwards the dentine cap is in direct continuity 

 with the basement membrane lying below the epidermis. Unlike 

 bone, the dentine grows on one sui^face only, and that is the 

 surface limiting the pulj)-cavity. The placoid scale, then, always 

 remains next to the epidermis, and never as a whole sinks down 

 into the dermis (text-fig. 196). As it grows older it may, and 

 generally does, acquire a well-marked basal jDlate of fixation 

 (text-fig. 196, b). This plate is merely an extension inwards of the 

 original cone, and it soon comes into connection with the deep 

 fibrous layer of the cutis. As described by Hertwig (6), the 

 connective-tissue fibres penetrate into the plate. Like the rest 

 of the denticle, the basal plate is never composed of true bone, 

 never contains bone-cells : it may be formed either of dentine, or 

 of some less dense substance especially in its deeper parts. When 

 the base is very large, as in the spines of Skates, the dentine may 

 gradually change to a softer tissue composed of numerous trabeculee 



